Each system here started with a problem a manager could describe in one sentence. Client and system names stay private. What each one does, and the problem it removes, are on this page.
The problem: Each site closed its day in a spreadsheet, and head office found the mistakes weeks later.
What it does: Each shift closes on one sheet: cash, card, credit sales and expenses. Audit accepts it or sends it back with the reason.
The problem: Cash left the site, and nobody could say when, or whether, all of it reached the bank.
What it does: Cash declared at each site is matched with what reached the bank. Unconfirmed deposits and repeated gaps are flagged.
The problem: Credit sales piled up with no clear limit for each business customer.
What it does: Credit sales are checked line by line, and each business customer gets a clear call: allow, limit, call or block.
The problem: Expenses and payments were approved in messages, with no record of who said yes.
What it does: Expenses, card and bank-transfer payments each have their own reviewer. A rejection goes back to the site manager with the reason.
The problem: Last month's figures kept changing after the books were closed.
What it does: Shift reports of a closed month are locked. Reopening one is a recorded step.
The problem: The finance director rebuilt the bank position by hand each morning.
What it does: Bank balances, money in and out, and bank reports in one read-only view for management and finance.
The problem: Promises to pay lived in phone calls, and nobody followed them up.
What it does: Who owes what, the promises to pay, and what actually reached the bank, in one desk.
The problem: Auditors spent their days ticking off matches that were obviously fine.
What it does: Rules clear the routine deposit and credit-sale matches, so auditors see the exceptions. Tested on past months against their own decisions.
The problem: Stock drifted between shifts, and nobody knew where the difference started.
What it does: Stock carries forward from shift to shift, and auditors confirm each day's goods in and out against the till system.
The problem: Shop counts lived in spreadsheets, and shortages were argued about for weeks.
What it does: Monthly shop counts entered site by site and supervised on screen. Surpluses and shortages are valued once audit approves.
The problem: Goods left the warehouse, and what arrived at the site was not compared.
What it does: Delivery notes for transfers are matched with what each site records as received, and differences are reported.
The problem: Management heard a delivery was late only when a site ran short.
What it does: The forwarder updates each load from a phone. Management follows its stage and gets an alert when it reaches the border.
The problem: Offers arrived by email and on paper, and comparing them meant retyping.
What it does: Offers and awards in one register, with the tender file built from the same form.
The problem: Orders were placed on instinct, so some sites ran short while others sat on stock.
What it does: Demand forecasts per site, checked against past weeks, with run-out risk and a seven-day what-if plan, as each site reports.
The problem: Payroll took days of spreadsheet work, with bonuses, penalties and advances tracked by hand.
What it does: Monthly pay worked out from the roster, bonuses, penalties and advances, with the bank payment list ready to send.
The problem: Contracts, leave and warnings sat in folders across the sites.
What it does: Staff records, contracts, leave, schedules and disciplinary letters, managed in one platform.
The problem: Management saw the numbers at month-end, when it was too late to act.
What it does: One view of sales, cash and exceptions, updated as each site reports.
The problem: Owners had to call the office to know where the money was.
What it does: Cash in the bank, sales and deposits on leadership's phones as each site reports, with an alert when a deposit looks unusual.
The problem: Site visits ended in a few photos and a promise to fix things.
What it does: Inspections on a phone: a checklist with photos, a rating, findings with owners and deadlines, and a report to management.
The problem: Contract and insurance deadlines depended on one person's memory.
What it does: Contracts, legal cases, insurance policies and deadlines in one register for the legal office.
The problem: Changing a price or posting a job needed a developer.
What it does: A company website the marketing team runs itself: locations, prices, offers and job posts, published without a developer.
The problem: A business association needed one public home, in several languages.
What it does: A multilingual public website for a business association, built and hosted by us.
The problem: Members could not easily find each other or compare offers.
What it does: Verified member profiles, requests for offers and offer comparison, as a clickable demo with the full build specified.
The problem: Supplier price rises slipped through, and shop margins shrank unnoticed.
What it does: Margins, supplier price rises and credit notes analysed from a verified copy of the records. Contracts and shelf photos are read on the device.
The problem: Reports needed someone to export data from the accounting system by hand.
What it does: A read-only copy of the accounting records that reporting tools and approved AI assistants can query without write access.
The problem: Sales, purchases, stock and ledger sat in separate screens of an old system.
What it does: One screen over the accounting records for sales, purchases, stock and ledger, with controlled document entry.
The problem: Inspection reports piled up faster than managers could review them.
What it does: Piloted: an AI reviewer checked submitted reports and photos against management's rules, logged each change under its own label and notified the inspector.
The problem: A simple question about the numbers took days and a specialist.
What it does: Ask in plain words. The assistant runs read-only queries, corrects its own errors and shows each query beside the answer.
The problem: Commitments and risks were scattered across mail, chats and notes.
What it does: Drafts the weekday briefing, tracks open commitments and risks, and prepares drafts that wait for the owner's approval.
The problem: Software changes reached users with mistakes nobody had caught.
What it does: One AI model drafts a change, a model from another provider reviews it, and a person approves before anything goes live.
The problem: The day started with someone rebuilding yesterday's numbers.
What it does: A daily AI-written finance summary, a question box and an on-request dashboard analyst, read-only by design. Ran from June to August 2026; now paused.
The problem: Sensitive figures could not be sent to an outside AI service.
What it does: An assistant on the office computer flags money and stock issues for staff to confirm and answers management's questions. Personal data is filtered out.
The problem: Checking drawings against office standards took senior engineers' time.
What it does: Checks sheets against office standards and compares revisions. Judgement and sign-off stay with the engineer; its AI step stays off until it is measured as reliable.
The problem: Nobody could say when the solar panels would pay for themselves.
What it does: Solar production and savings per site from inverter data, with payback once costs and tariffs are entered.
The problem: Contractors found out about overtime and idle machines at month-end.
What it does: Foremen log the day on a phone. The same day, the system flags overtime, slow progress, idle owned machines beside rentals, and missing reports.
A market-leading multi-site retail group, live across its network in four waves. Client named on request, under NDA.
Tell us which one. A partner replies, and we tell you what we would build first.
Tell us which one